European Commission logo
INSPIRE Community Forum

(How) do you include xxxDataset features in FeatureCollections?

This question is cross-cutting between different themes (e.g. LU, LC, SD, ...) that define a xxxDataset spatial object type that typically contains some metadata about the data set (e.g. spatial extent, validity period, reference documents, ...).

I would like to understand whether/how you make available this data set metadata when providing your data in a (wfs:)FeatureCollection, especially when providing them as a response to a WFS query. Will you always include the xxxDataset object in the response?

This is also related to the question about alternative encodings for coverages using "plain" coverages and moving the additional INSPIRE coverage attributes into the coverage metadata.

    • Roland GRILLMAYER

      By Roland GRILLMAYER

      In the WFS 2.0 Specification joined queries are foreseen to handle this issue ( WFS 2.0 conformance classes "Standard Joins", "Spatial Joins", and "Temporal Joins"). I had a discussion about this topic with Thorsten Reiz last week. He mentioned that one problem is that the specification of the joined query is not fully implemented in existing webservice technologies like Deegree and Geoserver. I have no practical experience on joined queries yet but will do some testing in the next weeks especially for the DS SpeciesDistribution | Land Cover | LandUse | Biotop and Habitat.

      There is a helpful discussion about the implementation of joined queries available in the degree github:
      For Geoserver see:

      The second problem we discuss for a long time with the domain experts on LC is, that the relation between the Dataset and the Dataunit is modelled in different ways. In most of the DS (Species Distribution | Habitat and Biotops | Land Use) the relation is established using an association which is traversable in both directions and encoded byReference (xlink:href). For Land Cover the relation is modelled as an aggregation – which ends in a different encoding (inline) and lead to a different handling of the WFS requests. Requesting a LandCoverDataset will response with one large LandCoverDataset instance in which all according LandCoverUnits are nested. I think this is not really a practicable approach. It is like downloading a predefined dataset from an AtomFeed – and that’s not the intention of providing data via WFS. A change request is submitted (i think was submitted by spain) to change the relation to an association and harmonise the DS.

      My pracitcal understanding to request WFS providing DS with a Dataset | Dataunit relation is that I first request all available Datasets. In a second request I select all DatasetUnits from the dataset of interest (for example article reporting 2013) using a filter on the association role referencing the specific dataset. To illustrate this approach and test results according the joined query to provide both feature types in one request I will come back as soon as testing is finished.

    Land Cover & Use

    Land Cover & Use

    Join this group to share your knowledge, learn and collaborate in solving issues related to the Land Cover and Land Use themes.